Critical Design Considerations for IoT Devices with Titu Botos

From the design level to the system level, connected products represent a new set of technologies and capabilities that can all serve as the weakest link in a product launch.

Our guest is Titu Botos, a PhD in electronics and mobile robotics and the CEO of NeuronicWorks, a product design and manufacturing company in Canada. 

We discuss how connectivity shapes both the challenges and the complexity of the product design process and where the most common gaps in design rear their ugly head. From radio compatibility, carrier compatibility, processor selection, and security we cover a lot of ground. This even includes the multiple factors that design decisions can impact battery life for IoT devices required to operate for long periods of time in the field without a power source.

This conversation makes for a helpful checklist for someone who is considering a product idea that they want to bring to market, how complex some of these decisions can be, and what the value of having an expert guide alongside them might look like.

Transcript

Ryan: 0:00

welcome to conversations in connectivity. I'm Ryan Carlson, your host. This is a podcast for the IOT professionals and product leaders responsible for building growing, executing, and at times educating others about the role that connectivity operations plays within their organization. You may not have the job title, but you've got the emotional scars to prove that you've been doing the job anyways. If learning how others are harnessing connectivity and the industries they serve, you're in luck.

0:26

Please consider subscribing and leaving a positive review if you find this episode valuable Today's guest is Titu Botos, A PhD in electronics and mobile robotics. And as the CEO of NeuronicWorks, a product design and manufacturing company in Canada. We discussed how connectivity shapes, both the challenges and the complexity of the product design process and where most gaps often rear their ugly heads in the design process. From radio compatibility, carrier compatibility, process, selection, and security. We covered tons of ground. It's clear though that from the design level to the system level, connected products represent a new set of technologies and capabilities that can all serve as the weakest I think you'll find that this conversation makes for a very helpful checklist for anybody considering a product idea that they'd want to bring to market. And it also serves as a reminder of how complex some of these decisions can be and what the value of having an expert guide alongside them might look like. This episode is brought to you by Soracom a global connectivity service provider that believes the fastest way to cost savings and scale is when customers are in full control of their connectivity operations experience, self service pay as you go global connectivity without a contract today at Soracom.io signing up for an operator account takes less than a So now onto the interview.

Ryan: 1:48

How long have I've been in Canada?

Titu: 1:50

I believe going on the 25th year now. Fantastic. And yes indeed, life is passing quite quick and the one thing that I am very happy about is that I don't look back one second. We are so happy to be here and even so proud to call us Canadian. We are, not that I necessarily leave my dream every day because yes, our life as a designer at the end of the day, even a manufacturer now has its challenges. However, yes, there are many things that we enjoy and be proud from.

Ryan: 2:25

I wanna pull on this thread just a little bit more of some reflection. t2, Boish, CEO at Neuro Works you. Have got a long CV with technical advisor, Chief hardware engineer, Vice President of engineering. But the thing that caught my eye was that at Polytechnic University, you've got a PhD with a doctoral thesis in 2003 for research regarding the adaptive control of autonomous mobile robots. Yes. A lot has changed since 2003. So what does T2 from 2003 think of the world of autonomous robots in 2022?

Titu: 3:04

Oh, that's a very loaded question. so 2003, the processors we had available at that time were, 2 86 until 2 86, 12 megahertz. And we are trying to, develop machine learning models on those processors. I remember that, having, Models of, neural networks of 100 or even 150 nodes neurons was recorded. So, Wow. definitely even at that time, we had the hopes up. We had the vision that this will come, yes. being younger, seeing movies you dream about, but I have to say that the reality in 2022 and soon enough, 2023 in many ways exceeds our expectations, our dreams of maybe how many 20 years ago, the capabilities, the power that machine learning can bring to systems in general, it's sometime I feel goosebumps and that much more because it is not. A dream. It is not a movie. It is a reality with age. I learn to enjoy, to appreciate little things. So what we are capable to do right now, the machine learning on a day-to-day applications, for me, it's very interesting, very rewarding, and actually a confirmation to what I dreamed before.

Ryan: 4:37

2002 Minority Report was out that Tom Cruise movie with the interfaces where he is moving, displays around with his hands when I bring up to people who know the movie and say, that was 2002. No way. Because we see that today the idea of, lidar and things that are built into the small little devices that are in our pocket, that we have the capability to do so much today, between, touring and metcalf's laws and all of the mm-hmm. the doublings of processors. it's interesting to be a native analog human living in a digital world, We grew up in an analog world. I remember modems dial up, in and getting clocks, even clocks and, how much things have changed. But even then, we're now living in a world in. even seven to eight years ago or so, we started seeing more and more cloud-based technology. we started seeing the early internet of things pre aws, I o t and Azure. And we're now in a world in which things connect to the internet on its own. It makes decisions on its own. Machine learning is a thing that we actually accept. There's entire jobs that exist for training machine learning models. There's machines that train machine learning models, right? Yes.

Titu: 5:59

Yes. it's crazy. It is, is crazy. So you are right. I was born and I was trained in the analog world. even my university days. Yes, we had some courses in microprocessors and all that, but still was a lot of analog. But part of my happiness, One thing that makes me happy, because I witness all this changes, and for a small part, we've been, as a company, part of this changes. I remember, I first started working in Canada, as a design engineer here. all the communication was wired, wired, wire wired, like RS 45, RS 2 32. can came after that. Ethernet came after that. But it was all wires, and then all a sudden wireless. Yeah, wireless and that, that came to be, And then first everybody was doing projects with Bluetooth. And on that note, I don't remember in the last seven, eight years to have a project that does not have some sort of connectivity. So it started with Bluetooth. Everybody was doing Bluetooth at that time. I believe because of the connectivity with the cell phones, the cell phone had a Bluetooth, so let's connect with something and use the cell phone to move data up to the cloud if necessary, if that was such a thing that there were only private clouds. So Bluetooth and then people moved into wifi designs, wifi. And for the last, I want there to say four to five years even, we moved into cellular. So moved the cell connectivity and that opens and yet another door, another, we are step referring to movies and games to another level of connectivity and that is great.

Ryan C: 7:26

I remember when. Wifi and Bluetooth were the predominant way of connecting devices wirelessly and cellular was still the only way you could get a cellular connection was to, buy a, like a hotspot and then you would. Low tech, you would end up, or you'd go through the insane cost of building out your own boards that had the ability to hold a sim, but the plans still weren't built around data. It was still, while you still had to provision the phone number, you weren't gonna be using any voice, you weren't gonna be using any text. and the data was an afterthought I had a trio 6 51 of the, those early keyboard phones when Black Barry was big. And you started thinking about like, how much data does my plan have?

Ryan: 8:16

they didn't. Yes, indeed. It was so little, I used two megabits last month. Oh my goodness. That's a lot of little text emails, right? so now we're in a world where you could go to a party and I've got a fun little party favors just, so let's try and name something that's not gonna be connected to the internet. And we can put our cynics hat on. And there's, that's a really hard question to answer.

Titu: 8:39

I truly believe there is no such a thing. Yep. Giving enough time. Everything will be connected. And I truly believe when I say that this iot wave, can we call it an iot wave? And then IOT has obviously a brother, a good brother called the in industrial iot. I iot. And so they go in parallel. I mention that because my background is industrial control and automation. so I believe that this wave of iot is not near the peak. we all have few good more years at decay, if not more of years until you see that, Yeah, we picked that wave. We have now more connected devices than not. There. I don't think there is anything there that should not be connected.

Ryan: 9:26

So with so many things that are being connected, I remember the reticence, even in the industrial internet things for people to connect things to the cloud. So first it was wireless is in, is not secure, right? there was always this fear. And now we've got, encryption, we've got protocols, we've got a variety of ways from the manufacturing level. There's now general acceptance of cloud use with private networks and virtual private networking. And, there's. We're now, if you were thinking about like the tipping point and that crossing the chasm on one side, you have all the early adopters where people were just looking at, Oh, this sounds neat and new. I want to be on the front end of that, versus crossing the other side of the chasm where it is just generally accepted, I believe, and have seen the conversation in the world. It's not should we connect our device, it's when do we do it and which one do we connect first? And then you get to go to how, and so that, that's where, neurotic works, chat steps in. So talk to me about the start of your neuronic work, neurons works journey, is, doing design, electronics, manufacturing. tell me about that.

Titu: 10:44

Where do I start? that's the big question.

Ryan: 10:46

what drove you to start it?

Titu: 10:47

I enjoy designing new things. And if you look at the quote unquote, the material of our people that are with us today, they enjoy that too. the power. You feel that when you create something out of nothing, when you put everything together and it starts to click, blink, buzz, whatever it does, it's a great feeling. it's a creation itself, as small as it is or as complicated it might be. It is something that was not there before and you made it. this is the primary driving force. Beyond that, on a second level, I can say that I. we try to, more than 13 years ago, we start, we tried to create a new product on the market at the, it was electronic shelf labels. We tried to marry together two technologies that were very new at the time. it was called Zero Power LCDs was the precursor of the E Inc. And then low Power RF that, the nine 15 megahertz proprietary protocols, communication. And that will help change this electronic shelf labels very quickly. And we came with a prototype and. Long story short, nine months later, ironically now my nine months later, we ran out of steam, as in two engineers, a prototype, no production, no, nothing. Basically trying to sell that to, to big store, big change stores like Metro and Dominion was at time and even Walmart and not getting traction. And then we look at each other. Now, what do we want to do? yeah, we love design, so let's compete, let's continue doing design. And slowly started by one, then two and three and more. More projects we've hired up to hire people. We created this business of design. We love design part. It is challenging, it is not easy, but has its own rewards now and then,

Ryan: 12:33

so the design process, I've seen it approached from multiple angles. Some people start with a significant business challenge and they work the word backwards asking how can technology meet the specific outcomes that we want? Now the question on the, like the grocery store or the retail shelving today, you see it all over the place. I've got a lot, There's a lot of grocery stores in my area where they automatically have the little updating tags and the cost of labor was getting higher and higher in certain markets. And eventually they said, How do we make this go quicker? But I've also seen people, the market doesn't always know what it really wants, and sometimes it takes pushing technology to create an example that can inspire people to go that, that, that might not be exactly what I want, but think of it in this other application. So talk to me about that. if we're talking about designing. Hardware solving problems. Every engineer I know, what drives them is the having the opportunity to solve really cool technology challenges or software design challenges or building something neat, right? It's not the money, it's not the fame. It's getting to solve that puzzle.

Titu: 13:59

Indeed. And even from that first lesson that we had more than 13 years ago, we learned a major lesson. We have to be in sync with the market. You have to recognize what the market, once you can be, it can be the coolest thing ever. If the market doesn't want it, you build it for the show only and you don't go farther than that. And I truly believe that 13 years ago or when we were ahead of the market, somehow, that lesson taught us that we have to talk with our client a lot, talk a lot with our client, understand him, try to understand the market behind him and design accordingly. But even though we do that on a daily basis, in fact, before we engage a new customer, we had about four, six even more where you just talk with them, try and understand what they want to build. But even if you do that, you have to build in order to build it. You have a one point jump in the cold water, take the chances, build something and bring it to life, and then present it to your customer, to your market. And again, receive feedback because Ryan, as you said, clearly many times people don't know what they want. They have to see it before they start to recognize the potential of the thing. So more, more, Often than not, you come with a prototype, come with a proof of concept on the market, then most viable product or little product and you just basically test the waters. The idea of, and what we see make our clients, some of our clients successfully is that they recognize this process. They embrace this process and they try to produce as they try to come on the market as soon as possible, test the water, so to speak Again, receive the feedback and rethink the opportunity. We've seen many times people coming to us with one idea. we produce an alpha, a better go on the market. They find that, for example, even a subsection of the of first idea turned around becomes a new product. They pivot, they go in a new directions and they become successful. And please let me know that this allow me to tell you they're not less happy just because of that. They become successful with completely new direction. Yes, we've seen that. And we are aware of that and we try to teach and tell our customer that, that way of working.

Ryan: 16:42

So what does that process, that consultative process look like when someone says, I have an idea, I wanna build this thing. Do you have a standard set of advice that you hand over or that you give to people? what advice do you give to people when they come with, they got money, they've got timeline, they've got intent, right? Like everything that would make a salesperson salivate. woohoo. But how do you pump the breaks to see that they're successful?

Titu: 17:08

Sometime we go directly and we present them our product development cycle. It's a cycle, by the way. It's a process, but in the end it's a cycle because it never stops. the holy grail of product design is that if you do it properly and you launch the product, you receive feedback and you go again for version two, version three, the whole process becomes a cycle, becomes a pinwheel that can maintain itself. So how do we go and we really open our, books, so to speak, teaching books, and we present to them the path for product development. And we show them how stages are, followed. One after the other. We try to explain to them that, So why we do that? Why we have the stages, why at each and after each and every stage you have a gate that somebody has to validate himself as an innovator or the market at large. So we have this process with stages, with gates, and what we try to achieve is to go for this fall often, fall quick, and pivot if necessary, such that they do not waste a lot of money. The worst thing one can do is to become enclosed sitting in between four walls with these ideas, pumping a lot of energy and resource into it to become a product on the market. Go there, I've done that, and go there and then there is no traction for. The best thing to do is to show results, involve people, suppliers, manufacturers, customers especially, and find feedback. Is it a viable, it's a viable product? It is an iterative process. I try, I'm trying to say here, and is something that we learned even the hard way for the last 13 years and more.

Ryan: 18:58

It's the most expensive way to prove your point. To go out and just build the thing, and sit in your own echo chamber. And, yes. I we've seen it over and over again. I remember working on a maritime that was a project for, that was a piece of hardware, connected hardware that would be used in maritime applications. And the initial ask that they came to us was saying, We wanna build an app. This is what we think people want. We want this app to perform these particular functions within a maritime vessel. and we said, That sounds great. How about we go and we talk to ship builders, ships captains, and some of the people within the channel that service these maritime vessels working on these specific aspects of a ship. And that took two weeks. And this is one of those things where you have to slow down to speed up. And the reports that came back is, there was not a single bit of feedback that validated anyone wanting to have an app. Nobody wanted it. But what we ended up walking away with was understanding where their real challengers were when they were coming into port, when they were out at sea and identified a number of things and ended up going a completely separate direction. We, we found out where their frustration came with and how to we, I think it ended up reducing an average of something like three dozen helicopter service trips. A year, three dozen and a helicopter trip to bring the parts that you need for a mission critical component within, a ship out at sea is a significant amount of money, especially when it's, under service contracts and where some of these people are under certain constraints. you never would find out how to make those leaps without stopping, pumping the brakes and asking why a lot. Right? We're seeing lots of connected products going out into the world, you're already getting people to think about that flywheel or the, that process that's not just a waterfall genius design. But creating that, that, feedback loop, we're now moving beyond people. Again, as we said earlier, asking why iot, Of course, we're gonna connect our thing. Now that people are actually connecting their stuff, they're even finding ways of monetizing their connected products. There's a whole new set of constraints to where companies are starting to internalize their ability to build and maintain these connected products. So if I said connectivity operations, what does that make you think of?

Titu: 21:46

That would be a bit of a long answer. So please bear with me here and please remember that everything I'm answering today is obviously from the perspective of a designer. Yes. So Internet of things and his brother, as I said, industrial iot has brought us into a connectivity, an age of connectivity that will enable machines and systems and objects to talk to each other, to function in new and expanded ways. We, we saw that, and that is the thing that brings us towards this new age. So we see now machines, objects communicate with each other, continuously forming large, integral systems, capable of creating and communicating and aggregating, analyzing and acting on data. The connectivity adds to design process an additional layer. Of complexity. And with that new set of challenges, actually we can say that the connectivity shapes the challenges and the complexity of the product design process. I see here four major, parts that transforms the new connected device. And that is, for example, the iot. The impact of the technology brings a new challenge, for example, marrying the physical li the physical, world with a digital world. Yes. And I said before, maybe this is the iot in itself, the connectivity. Therefore, in a way you can compare with the Pandora box. has a lot of advantages, but bring some challenges. there are many things to, to solve. how to incorporate an antenna in existing product, then security issues. So marrying a physical, object to the digital world, brings security issues. should I mention privacy? That's another discussion here. But yes, we are going there because there are advantages to have. this even now comes a new trend with this between, digital to the existing physical object. The other challenge is, always stay. You want devices to be always connected, to bring data, to report data or to receive, comments from the, from the cloud. So it has to be always on. And some of them are remote devices. They have to live on a battery. That's another set of challenges to have a long life on a set batteries. And we have designs that are, that are, leading based on our design data for 15 years on, on a battery. Again, it's obvious now that the other challenge is we move from a single system sitting somewhere in a quote unquote, in a corner of that system or of that object or of that machine being part of a huge, large system. Yes. right now the machine or the object does not have to behave. only its own, but has to behave according with entire, the entire system. That's another nice challenge.

Ryan C: 24:58

And it's not just that though, right? It's like that is a disparate data point that doesn't necessarily mean a whole lot in the absence of additional context. the temperature of something like, the temperature of a pump operating at a different elevation or, operating within a certain, climate. these things will have different implications or the vibration or, what is it connected to it? Is it moving water or is it moving tomato soup? you're really hitting a point that I think is that's part of that struggle that we're dealing with in the world of sensing is it's moving beyond just. Individual data points that we can monitor, but it is now making sense of the aggregate within context. we had a guest, he was talking about the digitalization of workforce knowledge, and it was the idea of the person who's been on the line for 40 years servicing a whole series of different pieces of equipment, understands a lot of that context and knows, Oh, what you're hearing is a bearing going bad whereas the sensor just knows it's getting warm or it's slowing down.

Ryan: 26:01

Indeed.

Titu: 26:02

and if I may have, I have one more point to make here. That is, is equally important, it's an opportunity and a challenge at the same time where this object or machine can continuously evolve in its uses because this connectivity brings us the opportunity to change, update in real time. Even the functionality of that, again, object machine system, whatever it is. So that gives us the opportunity to, make it better, so to speak, by changing the firm, changing the behavior of the machine based on the data that as a collectivity, the, the system brings data to the cloud and we analyze it and we make, we change the entire system with it.

Ryan: 26:46

I think that's probably one of the most important parts of recognizing the value of what is. Built into that world of connectivity is that things get better over time. Granted there's arguments about phones and planned obsolescence, that three year period, my phone actually gets better over time there's a lot of great examples of industrial Internet of things applications where designers like yourself built in pieces of hardware that weren't yet utilized. But knowing it's inexpensive to add it now Correct. Will add a dollar 37 to the build a materials knowing that we can extract hundreds of dollars a month worth of value just by making some remote firmware updates.

Titu: 27:33

Yes, indeed. And again, from a this designer perspective, talking about over the updates gives you a bit of peace of mind. Because you test a lot. Everything we design and produce, we design after the, at the end of the design cycle, we, sorry, we test, test, test, test. We manufacture, we test, test, test, test. And then you put it in the field and it fares not all the time, but it's a different world. But having this back door in a good sense, gives you the opportunity to update that firm, update that behavior, and gives you, a chance to make that design, that machine, bulletproof for the future.

Ryan C: 28:13

What I'm hearing is that building in the capability for remote access or a secure tunnel for making over the year updates Yeah. Gives you a lot more flexibility than you had previous to devices being connected

Titu: 28:28

Yes. That it is, and it is one of the, Crown jewels of the connectivity of this new age.

Ryan: 28:35

Amen to that. I, yes. Remember putting devices out into the car wash industry and before we had remote, had to like hand build remote access to tunnel in and to open up a session and flash firmware. Our number one cost that showed up on the overages on expenses was overnight. Indeed. And most people are like, what does overnight shipping have to do with connected product development? Like you put eight devices out that are running eight different car wash bays, and when one goes down because you're having an issue that you can't diagnose, you have to overnight a new replacement model. You pre-program it, put it in the box, send it out with another label, and they send you back the return so you can put it on the bench and figure out what went wrong.

Titu: 29:23

Yes, indeed. indeed. recalling units, it's one of the worst nightmares of a, of any company, let alone a startup. The moment you have a harder component of your product, having this, secure tunnel for firm updates can give you peace of mind. Recalling units can just. Shut off the company, basically. It's such a, such enormous cost for the company to bear that can shut you down.

Ryan: 29:49

If we're talking about companies, let's think about'em as an oem. They've worked with you to get a piece of hardware designed And if connectivity operations is their own understanding of how to deal with security, privacy, digital twin management, analyzing data remotely, accessing the devices, what gaps do you commonly see in OEMs that are newer to connectivity operations or to connect to devices

Titu: 30:18

indeed, we, I, we see quite a bit of gaps because connectivity and iot as a new technology, it comes with its own challenges. As I was saying before, IOT fundamentally has changed the dynamics of our product engineering. we have a new complexity to, to deal with, and not everybody is ready to deal with this new complexity. Speaking about gaps, I think we can talk at two different level. The first one is design. Very, very down design level, a harder level even. And the second level, it's up at the system. Now, when you talk about design, there are few things that people do not recognize immediately. For example, radio compatibility. Not everybody knows that in different regions of the world, you have different bandwidth of frequencies that one can use to connect the system, to the cloud. So you need a bit of expertise. You need to google quite a lot. You need to learn a lot to, to become the master of this table of frequencies and know what to use and where in fact you have to use a different modem for different regions of the world. And choosing because of this different in frequencies is not only the modem, but you have to change the antennas. You have to change the matching network in between the modem and the antennas. So there is quite a bit of deep knowledge, harder knowledge that one has to have in order successful design, connected product that will work in certain parts of the world or even around the world. Not to mention that we have a a design that works around the world. No matter where you are, you have to make some sacrifices in, in, in many respects from the engineering point of view. the next, there are a few points here. So first was the radio compatibility. Now the second one is carrier compatibility. Now knowing what SIMs to use what sim cards to use to make the connection again for different parts of the world. And I would like to take a moment here and, mention that. The eSIM, that, for example, Soracom offers, brings a huge advantage here. The same SIM card can be in insert in each device regardless of where it goes is going on in the world. As connectivity can be provided later, that delivers the single sku that is so important for operational and simplicity, particularly for large international iot deployments. This makes for simpler, more streamline manufacturing process and connectivity that uses local networks as local at local rates. So this is carrier compatibility it's another piece of information that is missing when somebody decides I want to create my own connected product. Third certification, and this is a topic in itself to certify a product, a wireless product. It's not for the faint of the heart or how they say it. it's a very. complex process that one has to master in order to pass each and every time. Then you have to add, you have to choose your processors. And today's, there are quite a few processors that one can use from. So use the proper processor for the proper application. It's another, skill. do I choose an overkill? Do I use one that cannot be used in two to five years or two or to five years? You have to have that knowledge to choose the right processor that will, provide the work done today, will do the work today, but also in five years from now maybe.

Ryan: 33:51

Isn't that interesting though that, So processors, I've learned a significant amount about the role that processors play, in iot, especially if, we're talking security and privacy, now we need to handle encryption. So now a processor, it's not just a matter of reading the sensor data or running any sort of onboard processes, we now need to handle encryption. And every time we make that processor more powerful, we need more power. And if we're using a battery, then it means it's gonna have a shorter battery life. Life, correct. And so it's this vicious cycle. I'm sure there's a great graphic out there of all of the trade offs. And so the design challenges I've seen is how do we offload some of those aspects? if we can, if the processor doesn't need to handle encryption, but if the SIM can handle encryption, Or we're already putting a sim in, it's already doing the encryption natively. That might mean smaller processor, lower cost, less power, consumed longer battery life.

Titu: 35:00

Yes. This, it's, this is one of the challenges of embedded devices. You have to achieve equilibrium between what is required, what is really required, and for example, the life of the battery. So selecting the processor and without selecting the operating system that runs on that processor is a crucial part. It's crucial element of designing a new connected device.

Ryan: 35:23

And the processor also will restrict, not a restrict, but it acts as a constraint on which protocols you can effectively leverage. Yes. MQTT or https Yes. Tls. what are the basics that people should understand about that processor selection process? Is there anything specific to connectivity that as general advice, do you say plan three, four years ahead if you want devices to get smarter over time?

Titu: 35:53

I don't think that anybody has, for example, an algorithm that one will, would punch in some parameters and get the name of a processor at the other end. That's fair. It comes, it comes with experience. What we're looking for, what they're truly looking for is, the usage of that processor today. Have I maximized the process of time? At the process we are doing today, we are gonna deploy first deploy the new unit. Yeah, the new connect connected devices. Have we, are we at 90% already? Or then it's a red flag, right? Right there and there. And with that comes again, operating system and the memory usage. are we, using them at a high level, 90% That is Right Flag. flag right then and then, it is hard to say that, Oh, I should make it a 20% or what is the optimal, load for the processor. But 90% for sure is not good. We have to understand the customer. You have to understand the market. You have to understand, what is the usage of that device. what we aim for. We aim for 15 years, then you better think twice about.

Ryan: 36:57

So you mentioned that there was hardware design gaps, and then there's system level design gaps. Did you wanna cover a couple of those? I'm curious to hear what those might be of. Was there, is there a leading gap at the system level design? Oh, yes.

Titu: 37:11

one that somehow always comes a surprise. And you touched already. It's security, security is, oh, we need security. Of course we need security. Okay. What can we do now? sure. this is one of the things that, comes as a surprise at 11 hour. And it should not because security to be com, to be fully implemented. it has to be implement across the board. So it's not only at the harder level. you have to have it in the algorithm communication algorithm. You have to have it in the cloud. It's. All over the system, but guess what? They want to implement security and the hardware is not capable of support. it might not have even the, because especially for, for example, monetary transactions, you have to have, part of the processor that addresses only, that, that encryption, the trust core of the processor. And you don't have it. You don't have it. You cannot implement it after you have to go for it. Then again, so security is one of the system level, it has to be applied that each and every level of the system.

Ryan: 38:10

I think that's the hardest part though. Like things that are gonna be, if it's gonna be a payment processing terminal, like in the state's, PCI-DSS compliant, the payment card industry data security standards for those that haven't built these systems in the past, the pci. same thing with medical, with a lot of the hipaa or even if there's things like high trust or SOC two or different targets that you shoot for. To, either pass an audit or clear some sort of check. I think what you've just uncovered though is absolutely true. It's security is an afterthought in industries that are not regulated where security represents. it's up to them to decide what level of risk is the customer willing to accept. And when you say, so is security optional, I think everyone say, of course not. It's not optional. Is security important? Yes, it's very important. And you go, So what level? what level of security are you committed to? The cheapest level possible? like that sounds.

Titu: 39:07

and the iron, The irony is that, for example, today, the technology evolving so fast, we are able to implement in our new designs, encryption and security much, much higher than a bank has today. That's very interesting because the banking industry is evolving very slowly. They have to have certain checks and balances before they go to the next best thing. But the nice part is you are able to implement something even more secure than that manage transaction. That said, as a, a, I want to mention that I thought that, it's hard to design and bring to market a medical product because of all the checks and balances and they have to do, it's nothing compared with a new, product that transactions, money from one side to the other, that is even harder. That, that's, that was very interesting to find out.

Ryan: 39:58

But you think about the level of, in the medical world, it is the a percentage of a fraction of a percentage of there being any sort of problem, especially if it's an implanted device or a device that's used in diagnostics. You're at risk of li there's a level of liability that could turn into a financial impact. Whereas in banking, we know for certain that crime happens. People are always looking for loopholes. it's, no, so it makes perfect sense that our financial, our electronic financial system must continue to evolve because crime is pushing. that is the external pressure at all times, is people are trying to exploit. System. So that level of vigilance makes perfect sense. So if security is an afterthought, talk to me about when things like connectivity are treated as the afterthought in the product development process.

Titu: 41:01

Yes. Besides security, one thing that we see again and again that people miss, at the beginning is, certifications. everybody believes, for example, we have this myth, and when believe that I'm using in my design, certified or pre-certified model, therefore my design is certified. Period. No, this is a myth and doesn't, it doesn't work like that. once you design the product, having a model inside, you have to recertify that the entire thing. That's the myth that we have to put out, forever. The other mistake that people do is here comes certification. They ghost full steam ahead. They do the alphabet, gama, whatever they go for pilot, and then they go for certification, which is very late. Why is that? Because they go for certification. They may not pass, especially if it's their first time trying to do that. And then guess what? They have? They are stopped on their tracks. They have go back to the almost square, uh, square one. They have to go to the, go back to the drawing board and redesign the. Almost from scratch. And that takes times and taxi resources, that is a very, bad mistake, to make. And that is one of the advantages working with companies like our, like us, like Neonic works cause we have years and years of experience designing your product and from the beginning, which path we take to make sure as much as it's possible, that it will pass with flying colors. To do that, we have a process that we, at every level of design we are talking to, we are testing as fast and as many times as possible. this new product in different forms such that we make sure that when we go for the final full certification we pass, we have connections and we know, What effects takes for FCC certification, for pt, crb, and even ul. So you have to have that knowledge, to be able to pass with flying colors from the first scope.

Ryan: 43:07

What I'm hearing you say sounds an awful lot. Like when I was in a software world, we had a fairly forward thinking QA department where they implemented a program where typically software would go all the way to the point where we're ready to send it off to qa and they would run it through all of the different, the scripts and the test scripts and looking for errors. And they instead took a QA lead and they called it integrated quality assurance. And on the project plan looked like we were slowing things down or adding to the budget. But by having someone that was even at the wire frame level going, Oh, you see what you're doing there, that's probably going to trigger a failure. We've seen this before. And so each. Small milestone in the process. They were actually having QA look over the shoulder and ask some of these questions even before there was enough code to write scripts against, and that ultimately made it so you didn't lose. All of that time at the end and found that it reduced the actual QA expense by upwards of 60% on a lot of projects.

Titu: 44:13

yes. This goes back to the idea of fail often, fail fast, and that implies bringing, talking and testing with the regulatory bodies of your system as many times as possible. And why is that? Because a change at the beginning of design cycle, it's way less expensive. The change at the end of the design cycle, as I mentioned at beginning here talking, you want to certify a pilot production already. That means you have a production. That means you, you spend a lot of money to bring that production up. If you have to change anything, it becomes unbearable, expensive. So yes, test as fast as many times as.

Ryan: 44:50

I'm hearing fail fast internally, don't fail fast externally.

Titu: 44:54

definitely. Yeah. that is the point. and I'm sorry it don't suggest anything else. but you have to do that the, by being scared of testing your product, it doesn't help you. You have to try to test it as many possible.

Ryan: 45:07

So I want to, turn to, a question that I pinned down. I love hearing what people have to say about this. So I believe that there is no such thing as a neutral brand interaction. So what, And what I mean by that is like they, you either, like you go to a restaurant and it's either a good or negative. They didn't fill my water glass up enough or, they weren't really nice to us when we came in, or Right. it's small, but there's, was just neutral. In regards to what people say about their wireless connectivity buying experience, would you say in general it's a positive or negative experience? when they're going and finding the carriers to support their products out in the world,

Titu: 45:53

it is hard to define the experience. And because of the complexity, because of the, of how many things one has to keep in his mind, his or her mind, for the successful product launch at the end of this effort, because there are so many things to think, so many ways the thing can go, go wrong. Unfortunately, buying connectivity, it's not an easy, pleasant, all time pleasant, Experience and it will sound self sufficient here or self inflicted or something with self here. But, working with companies like Soracom, like NeuronicWorks that we've been around the block few times, that know what is expected, that know how to, direct a customer or direct in the sense of guide a customer through this difficult process. It brings value to the table.

Ryan: 46:50

So outside of those experiences though, we went through where some of those gaps are. a company like NeuronicWorks, you bring to the table, Let's make sure we talk about certification, let's make sure we talk about antennas and processors and Right. You've got your checklist and it's not just pulling something off the shelf. for so many years, and this is personally. I'm new to the Soracom experience outside of Soracom. going to, big box carrier and just negotiating for cost and coverage. It felt so transactional. and there was so little value. It felt like I was insignificant because I wasn't moving a million devices, or I'm not buying, 10,000 lines of, of, of service. And I don't know that at this point, even 10,000 lines of service gets you more than just, you know, oh, you've got one person you can call. But do they have the answers? I'm not really sure. So, you know, for, for people that are, aren't familiar with the Soracoms and the neurotic works of the world, you've had to deal with a lot of these other, experiences out there. So what would someone at an OEM. What do you know of their experience? Where is it that they felt, underserved?

Titu: 48:03

I don't think I have a story per se here that I can share. Again, for me, it comes back to the challenge in between the, I want to go the market as quick as I can. but I don't know how to do that. And again, as you mentioned, Ryan, I like to start with a small sample set and therefore I have some thousands of units on the road. it is, I believe we, because we've been in their shoes, we are holding their hand along the way. It's sometime. It feels going up a mountain, it is that feeling where, it's not necessarily a walk in the park to, to make it happen. However, as I said, given the fact that you bring along the way, people that trust people that been there and done that, that can definitely make the experience, much easier and more pleasant.

Ryan: 49:01

Final question. So the doctoral thesis, from 2003. Yeah. What it, what it, what would, what would you hear in the future? What would you tell T2 of 2003 to look forward? Hm,

Titu: 49:21

more than you expect That is for sure. that was a thesis in basically machine learning that will guide a mobile robot to find its way around. yes, expect the unexpected thing that more will come, not necessarily the movie level at that level, but yes, it's coming step by step. And I see that coming and that makes me, even, more happier. I stayed in my mind with, with this love for, machine learnings called now then it was artificial neural networks and therefore even the name of the company. It's a, it's a, uh, of our company. It's a. Game of Wards, play of awards because neurons is that, that part of the brain that creates something new. It's a creation part. And the words is the something engineering, very regimented way to produce something. I'm happy. I'm happy where we are and I'm happy for we were going because I don't see any slowdown. On the contrary, there is a continuous acceleration of the, of new technologies coming on the market, and we see opportunities at each and every corner, not only for us, for our clients

Ryan: 50:22

as well. I wanna thank you so much for sharing your time, your expertise, and covering all of these gaps that I don't think we always think about. And so this makes a really helpful checklist for someone who is considering the idea that they may have, that they want to bring to market. And how complex some of these decisions can be and what the value of having an expert guide alongside them might look like. So t2, thank you so much. I appreciate everything. Is there any final advice you want to give to those of the future of the next generation 20 years from now?

Titu: 51:04

I like to steer the steal the awards and of many people and, something on the long along the lines of saying, the best is yet to come.

Ryan: 51:15

Couldn't agree more. Thank you so much, Titu.

Titu: 51:18

Thank you Ryan. Thank you for the opportunity. I hope that my views and my my words can make a difference and, um, we are here to help.

51:26

I got chills after that, I just finished listening to this interview all over again. And I have to say, I just love the fact that we have someone like T2, who is willing to share so much of their past experience. And didn't hold anything back. These are the, exactly the types of conversations that we're looking to have on this podcast. So please. reach out to teach you on LinkedIn and thank him for his participation. It would mean a lot. I appreciate it. All of our guests. This is exactly the types of thing that we're looking to do, especially for those of you that are new to IOT. Or currently in IOT and are looking for ways to look around the corner and see what it is that you're missing. Or how you can improve your own experience, how you can improve your career and what it is that we can all do together to raise the water level in this industry. Because there is certainly. The need. As professionals. To level, up to improve our game and find ways to overcome. What has been a lot of jobs just dropped in our laps. And. It's up to It's for those of us that have been here, done that, that we need to help. Make it better for the future generations of IOT professionals. Till next time. I'm Ryan Carlson. Take care.